第一部分：同种小鼠心脏移植模型构建及免疫机制研究

研究背景：

心力衰竭（Heart failure，HF）是所有心血管疾病终末期状态，部分患者经过规范化的药物或手术治疗可以缓解或治疗心力衰竭症状，但是仍存在部分患者尽管接受了规范的治疗后，HF症状仍然会持续进展表现为严重的心力衰竭症状与体征，即终末期HF。心脏移植（Heart transplantation，HTx）是目前治疗终末期HF最有效的治疗选择。尽管术后患者接受规范的免疫抑制剂治疗，但是术后3年内，仍有11%死于急性免疫排斥（Acute rejection，AR），关于AR过程中免疫机制仍有待探究。

 

研究方法：

本研究探究过程中，首先通过免疫组化探究人心脏移植AR患者组织中糖酵解水平。随后构建同种小鼠腹腔异位心脏移植模型用于移植排斥研究，包括同种同基因移植模型（Isograft，Iso）、同种异基因移植模型（Allograft，Allo）。随后对Iso供心、Allo及正常对照未移植心脏免疫细胞进行单细胞RNA测序（Single-cell RNA sequencing，scRNA-seq），对scRNA-seq数据进行降维定义分群、细胞比例确定、标记基因确定、RNA velocity、SCENIC分析、配体受体分析等。将分析发现使用对应小分子药物（LW6）在Allo模型中进行干预，并针对干预治疗实验再次进行scRNA-seq分析，干预治疗后的改变。同时再次利用临床心脏移植发生AR的患者进行多种染色验证发现。

 

研究结果：

一、心脏移植患者术后发生AR供心组织中巨噬细胞糖酵解水平增高

1. 本研究收集临床术后发生急性移植排斥患者心脏组织（n=3），同时病理上炎症浸润主要是CD3+T细胞及CD68+巨噬细胞。

2. 本研究采用多种标记染色及糖酵解中关键的限速酶—丙酮酸激酶（PKM）进行染色：发现PKM在AR患者中表达增高，尤其是巨噬细胞中。

二、构建小鼠腹腔异位心脏移植模型急性免疫排斥单细胞图谱

1. 本研究选择Balb/c小鼠作为供体，C57小鼠作为受体，采用腹腔移植的方法进行小鼠心脏移植。同时设置同种同基因组进行对照，其构建方式是Balb/c小鼠作为供体和受体，同样采用腹腔移植手术方式。并获取正常未移植组进行对照。上述3组各取5只小鼠进行实验，在第7天进行获取供体心脏进行实验。

2. 获取上述三组小鼠心脏分离细胞，流式分选Cd45+7AAD-细胞。使用10X Genomics平台进行单细胞建库，共获得46040个免疫细胞，其中Allo组为16972个细胞，Iso组为13480个细胞，Ctrl组有15588个细胞。

三、同种心脏移植免疫排斥心脏免疫微环境改变

1. 采用降维聚类分析共获得23群细胞，并根据特征性表达基因对23群免疫细胞进行定义。细胞数最多的细胞类型是单核-巨噬细胞（M，9群），随后是中性粒细胞（N，2群）、T细胞（T，3群），树突状细胞（DCs，3群）、B细胞（B，1群）、自然杀伤细胞（NK，1群）等。

2. 单核-巨噬细胞群共有9个亚群：M1是传统的单核细胞，M2细胞为中间态单核细胞，M3也表达Ly6c2和Ccr2，但是功能与M1完全不同；M4则是驻留巨噬细胞，M5特征性高表达Vsig4，M6高表达Spp1和Lgals3，M7高表达Fabp4，Gpihbp1，Sparc和Egfl7；M8表达抗炎相关的基因，M9为分裂期巨噬细胞。M1，M2和M3细胞比例在Allo组明显高于Iso组，CCR2是上述三者的共同标记分子，免疫荧光证实，CD68+CCR2+细胞在Allo组中细胞比例明显升高。M1细胞具有较强的抗原提呈能力与较高的糖酵解水平，并且糖酵解水平主要是在Allo组来源的巨噬细胞中水平高。M4-7主要存在于是正常心脏中，其中M4是正常心脏中最主要的免疫细胞类型（占所有免疫细胞比例58.8%）。M4与M5表达巨噬细胞M2极化相关基因，在Allo组逐渐被单核细胞来源的巨噬细胞所替代。M8在Iso中细胞比例最高，在Allo组中细胞比例低于Iso组但是稍高于正常对照，M8主要与血管生成与损伤愈合相关。

3. 4285个中性粒细胞可以分成2群。N1细胞在Iso组和Allo组中细胞比例相比Ctrl组增加，但是两组间未见明显差异。N1细胞高表达促炎性基因Il1b和Cxcl2。Allo组中性粒细胞与T细胞间相关作用关系更强：N1在Allo组中表达更多的趋化因子Cxcl10，该趋化因子与T细胞表面的Cxcr3结合激活T细胞。在Iso组，N1与T细胞间具有较弱的抗原提呈作用，而在Allo组中这种抗原提呈作用变得很强。

4. 2023个DCs，可以分成3群。DC1细胞比例在移植后降低并且在Allo组中细胞比例最低，DC2在Iso组中比例升高，而DC3在Allo组中细胞比例增高。DC3的标记分子为Ccr7，DC3向引流淋巴结移动，DC3还具有较强的吞噬作用及抗原提呈作用。

5. 2604个T细胞，可以分成3群。T1在Allo组中细胞比例增多，分成2个亚群：Cd8+T1细胞和Cd4+T1细胞。Cd8+T1细胞表达Xcl1与Ccl4，招募DCs细胞浸润。Cd8+T1表达Ldha，促进T细胞促炎表型；而Cd4+T1细胞高表达Hif1a和Got1基因，该基因能够调控T细胞增殖与T细胞的细胞毒性作用。Cd4+T1细胞还高表达Tnfrsf4，该基因是T细胞协同刺激激活分子。上述6个基因均在Allo组中表达高，同时Cd4+T1细胞具有调节T细胞激活和细胞粘附的作用。此外，Cd8+T1细胞主要具有细胞杀伤活性。T2细胞主要高表达干扰素刺激相关基因，如Ifitm3，Irf7和Gbp2；T2细胞主要存在于正常状态下；T3细胞特征性表达Trdv4，Tcrg-V6和Il17a，提示该群是γδ T细胞。

6. 1群NK细胞，共计1055个。在三组中NK细胞的细胞占比相似，约为2%。在Allo组，NK细胞表达更高水平的颗粒酶基因，如Gzma，Gzmb。Allo组NK细胞激活水平相比其他两组水平更高。基于SCENIC分析结果构建基因表达网络表明：Irf7在Allo组NK细胞激活过程中具有核心驱动作用。

四、Hif1α主要通过巨噬细胞调节急性移植排斥免疫微环境

1. 在Allo组中发现单核-巨噬细胞中糖酵解水平较高，而Iso组则没有增高。Hif1α在AR过程中主要的免疫细胞类型M1，M3和T1中表达水平较高，并且Hif1α在Allo组中表达水平较高。多种标记染色结果表明：HIF1A在Allo组表达增高，同时主要在巨噬细胞中表达增高。

2. 构建异种基因心脏移植模型，分成2组（未治疗组和LW6治疗组），给予治疗组腹腔注射HIF1A抑制剂LW6治疗，移植当天开始持续7天。病理结果表明，LW6可以有效地减轻Allo组中免疫反应程度（未治疗组vs. LW6治疗组，3.6 ± 0.3 vs. 2.6 ± 0.2，p=0.019）。同时LW6治疗组中Hif1α水平明显低于未治疗组（LW6治疗组vs.未治疗组，30.9 ± 13.5 vs. 191.7 ± 55.7，p=0.001）。

3. 本研究获取心脏移植AR患者心脏样本，并获取正常心脏组织样本进行对照。多种标记染色表明：AR中HIF1A表达高于正常对照患者，同时HIF1A主要定位于巨噬细胞。

4. 本研究对LW6治疗与未治疗小鼠进行免疫细胞单细胞测序。共获得31852个细胞，其中16197个细胞来自于LW6治疗组，15655个细胞来自非治疗组。分析结果表明：LW6治疗通过降低巨噬细胞抗原提呈作用及促炎能力减轻T细胞浸润，主要是Cd8+T细胞浸润。

 

结论：

本实验得到了以下结论：

确定探究同种心脏移植排斥的模型构建方法；

本研究借助scRNA-seq技术全面描述同种心脏移植AR过程中的免疫环境特征，为日后深入研究AR机制提供了细胞蓝图：

单核-巨噬细胞群：M1，M2和M3细胞比例在Allo组明显高于Iso组。M1细胞具有较强的抗原提呈能力与较高的糖酵解水平，并且糖酵解水平主要是在Allo组来源的巨噬细胞中水平高。M8在Iso中细胞比例最高，在Allo组中细胞比例低于Iso组但是稍高于正常对照，参与与血管生成与损伤愈合相关。

中性粒细胞：N1细胞高表达促炎性基因Il1b和Cxcl2。Allo组中性粒细胞与T细胞间相关作用关系更强。

树突状细胞：DC3在Allo组中细胞比例增高，标记分子为Ccr7，具有较强的吞噬作用及抗原提呈作用。

T细胞：T1在Allo组中细胞比例增多，分成2个亚群：Cd8+T1细胞和Cd4+T1细胞。Cd4+T1细胞具有调节T细胞激活和细胞粘附的作用，Cd8+T1细胞主要具有细胞杀伤活性。

NK细胞：在Allo组，NK细胞表达更高水平的颗粒酶基因，如Gzma，Gzmb。Allo组NK细胞激活水平相比其他两组水平更高。

本研究也发现了与AR相关的重要细胞类型（CD68+PKM+巨噬细胞）及基因-Hif1α，进一步强调巨噬细胞的价值，并发现Hif1α可以促进巨噬细胞糖酵解水平提高增加其抗原提呈作用与炎性招募作用，参与AR过程，同时靶向抑制Hif1α可以有效地减轻AR炎症反应。

 

关键字：小鼠异位心脏移植模型构建，心脏移植，急性免疫排斥，Hif1α，免疫代谢

 

 

第二部分：基于野生型供体猪构建异种心脏移植模型及免疫排斥机制探究

研究背景：

异种心脏移植是未来心脏移植发展的方向，尽管目前异种心脏移植取得了重大的突破，显著提高异种心脏移植后存活时间，但是关于异种移植过程中的免疫排斥机制仍不清楚。本研究将通过单细胞测序等先进技术深入探究异种心脏移植过程中最基本的免疫排斥类型，并结合单细胞测序数据评估目前的基因编辑策略是否可行。

 

研究方法：

本实验分别使用野生型巴马小型猪和成年的恒河猴作为供体和受体。采用腹腔异位移植的方法构建异位异种心脏移植模型开展实验，获取异种移植组供体心脏心室组织（n=2）与吻合血管（n=2），并同时取野生型猪心室及心脏血管作为对照，分离非心肌细胞进行单细胞测序及蛋白质组检测，并获取异种移植模型组外周血进行炎症因子检测。

 

研究结果：

一、构建野生型猪-非人灵长类腹腔异位心脏移植模型进行单细胞测序

1. 本研究选择国内特有的小型猪—巴马小型猪作为供体，选择恒河猴作为异种心脏移植受体。

2. 将野生型猪心（n=2）异位移植到猴腹腔中，第一只存活约30分钟。第二只存活时间约为2小时。将吻合口血管（主动脉及肺动脉）及心室（左、右心室）分别取下来分别进行酶消化和单细胞建库。另外取两只同窝野生型猪作为对照，最后野生型对照心室及血管各2个文库，移植组心室及吻合血管各2个文库。

3. 共获得86708个细胞，其中移植组细胞数为47484，对照组为39224。对照组所有文库细胞均来自猪；而移植组4个文库的猪猴细胞比例分别是：第一只心脏猪细胞比例为95.7%，猴细胞比例是4.7%；第一只血管猪细胞比例97.9%，猴细胞比例2.1%。第二只心脏猪细胞比例为3.0%，猴细胞比例97.0%；第二只血管猪细胞比例为99.8%，猴细胞比例是0.2%。猪细胞数为85631个，猴细胞为1077个。

二、超急性免疫排斥供体细胞改变情况

1. 75510个非免疫细胞（>90%）分成8群：成纤维细胞（Fibroblast，FB）、平滑肌细胞（Smooth muscle cells，SMCs）、内皮细胞（Endothelial cells，ECs）、神经元细胞（Neuron）、淋巴内皮细胞（Lymphatic endothelial cell，LEC）和分裂细胞（Diving cells）和心肌细胞（Cardiomyocytes，CMs）。非免疫细胞中细胞数最多的细胞为FB，随后依次是SMCs、ECs、神经元细胞1群、LEC、分裂细胞、CM及神经元细胞2群。

（1）成纤维细胞参与抗原提呈及血栓形成过程：FB可分成 6群。FB3与FB4在异种移植组中细胞比例显著提升。FB3为肌成纤维细胞，功能是抗原提呈和肌肉收缩。FB3在异种心脏移植中表达SLA-DRA，发挥抗原提呈作用。FB4表达SERINE1参与血栓形成，同时表达趋化因子CCL2。即，肌成纤维细胞发挥抗原提呈作用，SERINE1+成纤维细胞参与异种移植血栓形成。

（2）平滑肌细胞在异种移植过程中发挥免疫调节的作用：SMC分成5群，1群是血管特异性平滑肌细胞，SMC3群细胞比例在异种移植心脏中增高。SMC3群主要功能包括：调节免疫反应和细胞增殖。

（3）分泌型内皮细胞在异种移植过程中增多：EC可以分成4群，1群在异种移植过程中细胞比例在心肌和血管中均增高。1群细胞（MT1X+内皮）与多肽合成及酰胺合成通路有关，为分泌型内皮细胞，参与移植物衰败。

2. 猪免疫细胞异种移植后改变：共8256个猪免疫细胞，可分成13个群：巨噬细胞（2群）、CD8+T细胞、NK细胞、CD4+T细胞、单核细胞、分裂的CD8+T细胞、树突状细胞（2群）、分裂的巨噬细胞、B细胞、肥大细胞和浆细胞。巨噬细胞是猪免疫细胞中数量最大的一类细胞、随后是CD8+T细胞、NK细胞等。CD8+T细胞和NK细胞比例在异种移植心脏中细胞比例明显低于对照组比例，而巨噬细胞1群（HMOX1+巨噬细胞）在异种移植过程中细胞比例增高，主要发挥防御反应。CD8+T细胞与NK细胞具有较高的迁移活性及细胞趋化作用。

三、受体免疫细胞浸入至供体心脏参与异种移植免疫排斥反应

1. 收集异种心脏移植模型不同阶段外周血，包括的时间点有：术前、吻合后5分钟、吻合后10分钟、吻合后20分钟、吻合后30分钟、吻合后1小时、吻合后3小时等。使用非人灵长类37因子试剂盒进行检测外周血样本，趋化因子水平在异种移植后持续升高，如MCP-1、MIP-1β、Eotaxin与MIG。

2. 1072个猴细胞，分成11群：CD8+T细胞、单核细胞、CD4+T细胞、NK细胞、FB、B细胞、巨噬细胞、SMC、EC、血小板以及DC。CD8+T细胞、单核细胞以及CD4+T细胞是数量最多的细胞类型，NK细胞浸润细胞细胞数约为上述细胞类型的1/3。CD8+T细胞主要功能是细胞毒性作用；CD4+T细胞可能在多肽合成中起到辅助作用；单核细胞的功能可能是迁移至供体心脏中促进炎症反应。

3. 猪内皮表面CAV1受体与猴树突状细胞、猴CD8+T细胞、猴CD4+T细胞等细胞相互作用，提示CAV1是异种免疫排斥过程中最重要的受体。

四、野生型猪作为供体引起超急性免疫排斥的蛋白水平改变

1. 本研究对异种移植心脏（n=2）与野生型心脏（n=2）进行TMT蛋白质检测，技术重复一次。将|Log₂FC|≥1且FDR<0.05定义为差异蛋白，异种移植组表达上调的蛋白有39个，而下调的蛋白有32个。

2. 本研究差异基因主要属于如下几类：（1）核糖体构成相关蛋白，如RPL34、RPL3、RPL14等；（2）参与T细胞反应的蛋白：LOC100515788、CYFIP1、IGSF8、EPB41、ROCK2；（3）参与血栓形成的蛋白：THBS1、HMGB1、ROCK2等；（4）潜在猴来源的蛋白：C4A、IGSF8、SRSF7、DDI2、EPB41、SNRPG，猴来源蛋白主要参与调节补体及T细胞反应相关通路。差异蛋白相互作用网络主要是：中性粒细胞与血栓形成通路、IL-2产生通路及RNA剪切通路。

五、新型供体猪基因编辑策略推测

1. GGTA1基因表达比例最高为16.35%，随后是CMAH（3.20%）和B4GALNT2（0.13%）；成纤维细胞、平滑肌细胞、内皮细胞和心肌细胞是表达异种抗原数量最多的非免疫细胞类型。其中，内皮细胞是表达异种抗原比例最高的非免疫细胞，比例为36.5%。在免疫细胞中，巨噬细胞、树突状细胞、B细胞和浆细胞高表达上述异种抗原：18.8%~58%。分裂细胞通常比其他细胞表达上述抗原比例更高。

2. 目前对猪进行基因敲除或修饰涉及9种不同基因转入，包括：补体调节蛋白（CD46、CD55、CD59）、抑制NK细胞功能（B2M、HLA-E）、抑制巨噬细胞功能（CD47）、血栓调节蛋白（hTHBD、hTFPI及hCD39）。

（1）调节单核-巨噬细胞活性：单核-巨噬细胞抑制性中，发现SIRPα在异种移植组表达较低，此外，除了该信号通路以外，PDCD1及CD155介导的抑制功能受体表达量也较低。激活性受体LGALS3、FCAR及ITGAX在异种移植过程中表达水平较高。

（2）调节NK细胞在异种移植过程中的直接杀伤作用：NK细胞表明抑制型受体KLRB1与KLRD1表达水平相似。此外，NK细胞表面激活性受体：ITGB2、CRTAM、KLRF1、NKG2D、SLAMF7及CD226在异种移植NK细胞中表达增高。

 

结论：

本研究为异种心脏移植模型提供了国内适用的构建方法；

利用单细胞技术及蛋白质组数据揭示野生型供体猪异种心脏移植免疫排斥反应发生过程中猪猴细胞改变情况：

猪非免疫细胞改变：心脏中主要的非免疫细胞是成纤维细胞，肌成纤维细胞发挥抗原提呈作用，SERINE1+成纤维细胞参与异种移植血栓形成过程；LPL+SMC在异种移植中比例增加，通过调节免疫反应和细胞增殖参与异种移植免疫排斥；内皮细胞1群（MT1X+内皮）为分泌型内皮细胞参与移植物衰败；

猪免疫细胞改变：CD8+T细胞和NK细胞比例在异种移植心脏中细胞比例明显低于对照组比例，而巨噬细胞1群（HMOX1+巨噬细胞）在异种移植过程中细胞比例增高；

外周血炎性因子及猴浸润细胞：MCP-1、MIP-1β、Eotaxin与MIG趋化因子水平在异种移植后持续升高，招募T细胞及单核巨噬细胞浸润供心参与异种移植超急性免疫排斥过程中杀伤供体细胞；猪内皮上受体CAV1可与猴免疫细胞的多种配体结合，提示受体CAV1可能是异种免疫排斥过程中最重要的受体；

蛋白水平改变：差异蛋白相互作用网络主要是：中性粒细胞与血栓形成通路、IL-2产生通路及RNA剪切通路；

基于异种移植免疫排斥机制对现在基因修饰方案进行评估，为构建新型基因编辑供体猪提供新的理论依据：a. 异种抗原敲除：目前敲除三种异种抗原，GGTA1，CMAH和B4GALNT2；b. 转入基因评价：关于抑制单核-巨噬细胞上，目前转入基因有HLA-E及B2M，改为HLA-E+CLEC2D；抑制NK细胞功能上，将CD47改为TIGIT；c. 补体调节蛋白（CD46、CD55、CD59）和血栓调节蛋白（hTHBD、hTFPI及hCD39）保持不变。

 

关键字：异种心脏移植，单细胞测序，野生型供体猪，蛋白质组，新型基因编辑猪

 

 

第三部分：单细胞测序揭示自身免疫心肌炎免疫环境动态改变

研究背景：

心肌炎（Myocarditis，MCD）是一种以炎症浸润及心肌损伤为特征的心脏疾病，可以进展成扩张型心肌病（Dilated cardiomyopathy，DCM），严重进展患者需要接受心脏移植（Heart transplantation，HTx）治疗。异常的免疫反应是导致临床症状及病理特征的根源。心肌炎根据病理表型及疾病进程可以分成3个阶段：急性炎症期、亚急性炎症期及慢性心肌病阶段。心肌炎炎症浸润细胞复杂性目前仍未解决。本研究旨在探究自身免疫心肌炎从早期炎症转变为心肌病过程中不同阶段的免疫系统变化，确定与炎症反应程度相关的基因。

研究方法：

本研究使用心肌球蛋白α重链诱导自身免疫性心肌炎（Experimental autoimmune myocarditis，EAM），构建不同阶段自身免疫心肌炎，包括正常对照、急性炎症期、亚急性炎症期与慢性心肌病阶段。通过酶消化法消化分解心肌组织，流式分选心肌组织中Cd45+细胞，将获得的免疫细胞进行单细胞RNA测序（Single-cell RNA sequencing，scRNA-seq）。随后对获得的单细胞数据进行降维分群、定义细胞群、确定标记基因、不同细胞功能及重要的激活转录因子。并对不同免疫细胞亚型在不同阶段的比例进行分析，同时分析炎症基因与纤维化基因在不同细胞中的比例情况，确定与炎症相关的基因。将获得的发现在EAM模型进行干预探究，最后利用心脏移植心肌炎心脏标本进行验证。

研究结果：

一、EAM模型免疫微环境动态单细胞测序

1. 经过对不同剂量下EAM进行病理表型鉴定发现，250μg条件下，病理表型最明显。对于EAM构建前通过超声心动图进行筛选将心功能异常的小鼠排除在建模过程。构建不同阶段EAM模型用于后续实验中，超声结果初步提示：不同阶段EAM模型具有典型的超声特征。不同阶段EAM的炎症细胞浸润情况不同，14天炎症浸润最为显著，60天炎症浸润明显减弱。

2. 对正常对照、急性期、亚急性期及慢性心肌病期阶段，使用Cd45抗体进行流式分选心脏心室肌组织中免疫细胞，采用5端建库方式进行scRNA-seq，共计有34665个细胞纳入分析，其中来自对照组（0天）有6888个（19.9%），急性期（14天）10351个（29.9%），亚急性期（21天）10821个（31.2%）及心肌病期（60天）6605个（19.1%）。

3. 初步分群可以获得26个细胞群，定义成9个细胞大群，包括：巨噬细胞（M，Adgre1）、中性粒细胞（N，S100a8）、T细胞（T，Cd3g）、自然杀伤细胞（NK，Ncr1）、B细胞（B，Cd79a）、树突状细胞（DC，Flt3）、自然淋巴细胞（ILCs，Gata3）、心肌细胞（CM，Actc1）和内皮细胞（EC，Cd34）。不同阶段免疫细胞构成表明：（1）巨噬细胞是心脏主要的免疫细胞类型，不管是正常状态还是疾病过程，巨噬细胞比例均＞60%；（2）正常状态下，中性粒细胞比例低于2%，炎症发生后中性粒细胞比例急剧增高（14天，18.79%），且中性粒细胞在急性炎症期所占细胞比例最高；（3）正常情况下，T细胞占所有细胞比例为7.35%，炎症急性阶段T细胞所占比例为9.86%，在亚急性阶段中T细胞比例为14.94%，此时T细胞比例达到最高，随后T细胞比例降低，在心肌病阶段T细胞比例为10.01%。

二、不同免疫细胞类型在EAM发生发展中的作用

1. 22348个巨噬细胞，分成9个亚群。M2的标记基因是Cxcl9，M3的标记基因是Ccl8，M8的标记基因是Vsig4，以及M9的标记基因是Treml4。M2，M3细胞比例变化与EAM炎症程度相一致，并且M2是急性炎症期及亚急性炎症阶段中主要的免疫细胞类型，而M1，M8与炎症程度趋势相反。M2细胞特征性高表达Nos2，Arg1和Ass1基因，该群细胞主要的功能包括一氧化氮（NO）生成过程、吞噬体、通过MHCⅠ抗原提呈作用等。Ccl8在M3中高表达，同时该群细胞是亚急性阶段中主要的细胞类型，且该群在亚急性炎症阶段中细胞比例最高。M8高表达Tnf，Il10和Vsig4，且该细胞比例在14天最低，随后随着EAM进展细胞比例逐渐升高。M8的功能包括正向调节防御功能。M7高表达M2极化相关基因Clec10a及Tnip3，且该群细胞比例在60天处于最高水平，M7细胞功能是抑制心肌病阶段免疫反应。转录因子Bhlhe40，Nfil3，Mtf1，Esrra，Rnux3和Hif1a被预测是M2的潜在激活转录因子。对于M3，潜在激活的转录因子有Crem，Irf7，Bach2，Hif1a和Rxra。

2. 4299个中性粒细胞，其中超过86%的中性粒细胞来自炎症阶段，可以被分成3群。发现Il-1a在N1和N3中表达升高，而Il-1b在三个分群中表达均增高尤其是N1。N1是14天中主要的中性粒细胞类型，N1具有分泌细胞因子的功能，如IL-6和TNF、防御反应、中性粒细胞脱颗粒以及炎症反应，N1（Il-1a⁺中性粒细胞）具有促炎的作用。N2是心肌病阶段主要的中性粒细胞成分（超过80%），同时高表达促纤维化基因Mmp8。N2富集功能包括：正调节MAPK级联信号通路和肌动蛋白细胞骨架组织。N2（Ngp⁺Mmp8⁺中性粒细胞）是参与EAM心肌病演变的主要中性粒细胞类型。

3. 3805个T细胞，分成5群。T2的标记基因为Tnfsf8和 Cxcr6，也高标达Cd4、Il-17a、Rora 和Tnfsf11，T细胞2群为Th17细胞，在14天比例最高，并且该细胞群细胞比例变化趋势与心肌炎炎症趋势一致。T4细胞高表达Il23r、Il-17a、Tcg-v2，因此该群细胞为γδ T细胞，T5细胞也是γδ T细胞。T4细胞表达Il-17a，定义为Il-17a⁺γδ T细胞，而T5细胞不表达Il-17a被定义为Il-17a^-γδ T细胞。T4细胞为60天中主要的T细胞类型。TCR测序发现：TCR克隆主要存在于21天的Th17细胞。Th17特征性转录因子（Hif1a、Rora 和 Nfkb1）及Xbp1以及其下游基因在T2细胞（Th17细胞）中均高表达；Rara、Atf6、 Fosb和Rora.1及其下游基因，且Rarg调节的凋亡相关基因在T4细胞（Il-17a⁺ γδ T细胞）中高表达。综上，Il-17a⁺T细胞在心肌炎炎症启动及心肌病转化过程均发挥作用：Th17细胞在炎症早期阶段起到启动炎症的作用及Il-17a⁺γδ T细胞则促进心肌炎向心肌病转变形成。

4. B细胞1459个，可以分成2群。B1高表达Ighd，然而B2高表达Igkc，Iglc1，Iglc2，Iglc3和Ighj3。B1高表达Ms4a1与Cd19，B1是所有阶段中主要的B细胞类型，占比在92.2-98.5%。BCR测序结果提示：BCR克隆主要集中在60天阶段中B1，说明该群在心肌病阶段BCR克隆增多。B2特异性高表达Jchain和Eaf2，B2细胞比例随着疾病病程逐渐增高，在心肌病阶段中细胞比例最高。B1细胞中由转录因子Elk3、Max、Irf2、Ikzf1、Spib、Pax5 和Sox5调控的基因表达水平高。浆细胞特征转录子，如Prdm1、Atf6、Xbp1和Creb3，在B2中转录活性增高，同时这些转录因子及其靶基因在B2细胞中均处于表达高水平，B2细胞是浆细胞，可能通过产生抗体参与EAM过程中自身免疫反应。

5. 1077个DC细胞，可以分成3群。DC1细胞高表达Cst3和Sept3；DC2细胞高表达Ccr9和Klk1；DC3细胞高表达Il-12b和Cacnb3。DC3细胞在正常状态下所占比例极低（<1‰），其细胞比例随着EAM发生而增高随后逐渐降低。DC3细胞主要参与抗原处理及提呈、髓系细胞分化和产生分泌细胞因子，如TNF。由Runx1与Irf5调节的基因在DC1细胞中高表达。由Ehf、Nfkb1、Stat1、Rel、Nfkb2和Stat4调节的基因在DC3细胞中表达增高，Rel（NF-κB的亚单位）可以促进DC3细胞中表达Il-23。DC3细胞可能通过细胞内NF-κB激活释放Il-23调节Th17细胞反应参与心肌炎炎症反应过程中；此外，T4细胞表达Il-23r，本研究DC3细胞还可能参与T4细胞（Il-17a⁺γδT细胞）激活过程中。

三、靶向Hif1α可以有效减轻自身免疫心肌炎炎症反应

1. 通过比较发现，M2细胞与Th17细胞这两种细胞类型中炎症基因表达量及表达比例是14天阶段中最明显的。Hif1a是M2细胞与Th17细胞两种细胞类型共有的激活转录因子。在26个细胞分群中，细胞中Hif1a表达水平与细胞群的炎症评分成正相关（Pearson’s r=0.69，p=0.001）。Hif1a表达水平在14天最高，随后逐渐降低，表明Hif1a可能通过激活急性期相关细胞群参与EAM的发病机制中。

2. 本研究使用小分子PX-478探究抑制Hif1α对心肌炎是否具有治疗作用，将构建成功的EAM模型分成3组：急性期（14天）、亚急性期（21天）及心肌病期（60天）。再对分组中的一半小鼠于取材前1周连续腹腔注射PX-478进行治疗，另外一半小鼠腹腔注射等量生理盐水作为对照，并在取材点进行取材和表型鉴定：（1）流式结果不同阶段EAM免疫细胞浸润比例较高，治疗组Cd45+细胞浸润比例显著低于EAM组，与对照组Cd45+细胞比例相似；（2）在EAM不同阶段可见炎性细胞广泛浸润，治疗组与对照组均未见明显免疫细胞浸润。IHC结果表明：14天治疗组Hif1a表达水平明显低于14天急性炎症组并且与对照组表达水平相似；14天治疗组CD3及CD68表达水平明显低于14天未治疗组。

3. 本研究收集5例自身免疫心肌炎移植患者、5例慢性心衰患者移植患者（选择DCM患者）及5例正常供心心脏样本，结果表明：与DCM和正常对照相比，自身免疫心肌炎心脏组织中单位mm²内HIF1A+细胞数较高（分别是4106.0 ± 423.6 vs. 605.6 ± 213.7 /mm²，p<0.0001；4106.0 ± 423.6 vs. 821.6 ± 223.0 /mm², p<0.0001）。最后通过多种标记技术探究自身免疫心肌炎患者中HIF1A+细胞类型，结果表明主要是T细胞、巨噬细胞及少量的心肌细胞表达HIF1A。上述结果表明：自身免疫心肌炎患者心脏组织中也高表达HIF1A，可以在未来用于自身免疫心肌炎患者治疗靶点。

 

结论：

单细胞免疫细胞测序描述了不同时间点EAM动态演变过程中心脏免疫微环境变化情况，揭示了不同阶段中EAM心脏免疫特征，使得人们更加深入地理解EAM免疫系统演变过程。通过单细胞测序工作，本研究确定了EAM中与炎症反应强弱相关的细胞亚群。此外，确定并验证了Hif1a可以作为治疗自身免疫心肌炎的潜在靶点。

 

关键字：自身免疫性心肌炎，免疫微环境，单细胞测序，Hif1a，PX-478

Part 1. Model of allogeneic heart transplantation and immunological rejection

Background:

Heart failure (HF) was the end stage of almost all cardiovascular disease, some patients still developed into end-stage HF with standard drug treatment. Heart transplantation (HTx) remains the ultimate treatment option for patients with end-stage HF, which is defined as the presence of progressive and/or persistent severe signs and symptoms of heart failure despite optimized medical, surgical, and device therapy. In patients who undergo HTx, acute rejection (AR) accounts for about 11% of deaths in the first 3 years. Moreover, recurrent AR has a cumulative immune injury effect on the onset of cardiac allograft vasculopathy, which is an important contributor to graft failure. AR is an important contributor to graft failure, which remains a leading cause of death after HTx. The precise immunological network of AR remains unknown.

 

Methods:

The glucose metabolism of immune cells in human heart tissues with AR was investigated by immunostaining. To find new therapeutic targets, the mechanism of AR was further explored by single-cell RNA sequencing (scRNA-seq) analysis of Cd45+ immune cells extracted from isografts, allografts, and untransplant donor hearts.

Results:

Glucose metabolism of macrophages elevated in patients with AR

Explanted hearts (n=3) were collected from patients receiving a second heart HTx due to AR. Besides T cells (CD3⁺), macrophages (CD68⁺) were found in the explanted hearts with AR. Pyruvate kinase (PKM) was highly expressed in AR, especially in macrophages. This suggested that the glucose utilization of graft-infiltrating macrophages is higher in AR.

Total cardiac immune cell populations among three groups

For searching potential therapeutic targets, scRNA-seq was performed on the total cardiac immune cell population (CD45⁺) from allografts, isografts, and untransplanted hearts. According to the histological analysis, AR only happened in allografts, with obvious inflammatory cell infiltration and myocardial cell necrosis. Transcriptional profiles of 46,040 cells were captured after quality control (allografts: 16,972; isografts: 13,480; control: 15,588). 23 cell clusters were identified according to the well-characterized marker genes and the enrichment function of each cell type. The most abundant cell populations were monocytes/macrophages (M, 9 clusters), followed by neutrophils (N, 2 clusters), T cells (3 clusters), dendritic cells (DCs, 3 clusters), B cells (1 cluster), natural killer (NK) cells (1 cluster), innate lymphoid cells (ILCs, 1 cluster), cardiomyocytes (1 cluster), fibroblasts (1 cluster), and contaminated cells (1 cluster).

The phenotypic shift of monocytes/macrophages in AR

We defined 9 transcriptional states of monocytes/macrophages. The gene expression pattern was different among monocytes/macrophages subclusters. M1 indicated these clusters belonged to classical monocytes. M2 could be defined as intermediate monocytes. M3 also expressed Ly6c2 and Ccr2 but had completely different functions compared to M1, M3 is mainly associated with inflammatory response, including response to IFN-β (interferon-beta), leukocyte chemotaxis, and regulation of innate immune response. M4 highly expressed markers of resident macrophages. M5 expressed Vsig4. M6 expressed Spp1 and Lgals3. M7 expressed Fabp4, Gpihbp1, Sparc, and Egfl7. M8 expressed the marker genes associated with anti-inflammation, including Fn1, Arg1, Lrg1, and Olr1. M9 represented the dividing cell, which had the highest cell-cycle score. The proportion of M1, M2, and M3 cells all increased in allografts compared with that in isografts. CCR2⁺ and CD68⁺ monocytes/macrophages were confirmed by immunofluorescence to be increased in allografts. M1 may have a strong antigen-presenting ability, and maintained a high level of glycolysis. M4/5/6/7 were mainly found in normal hearts, they had low expression of Ly6c2 and Ccr2. Among them, M4 was the largest number of immune cells in normal hearts. Enrichment analysis showed M4 was associated with ERK cascade and regulation of hemopoiesis. M4 and M5 expressed genes associated with M2-like macrophages (Mrc1, Maf, Cbr2). These two clusters of macrophages were substantially replaced by monocyte-derived macrophages in allografts. M8 was observed mainly in the isograft, but also showed a slight increase in allografts. M8 was associated with angiogenesis and wound healing.

The pro-inflammatory role of neutrophils in AR

We detected 4,285 neutrophils, which were divided into 2 clusters. The proportion of N1 was significantly increased in both Iso and Allo groups, but no difference was found between the two groups.N1 released high levels of proinflammatory genes Il1b and Cxcl2. There was stronger interaction between neutrophils and T cells in allografts. N1 in allografts expressed more Cxcl10, interacted with CXCR3 on T cells. N1 in isografts had a weak and sparse antigen-presenting relationship with T cells (H2-Q10/Cd3), but the relationship in allografts was stronger. This suggested that neutrophils play different roles in AR and ischemia-reperfusion injury. In AR, neutrophils had more pro-inflammatory effects.

Ccr7⁺ Dendritic cells were responsible for antigen presentation

2,023 DCs were detected in our study and identified as 3 types of DCs. The percentage of D1 decreased after HTx, the percentage of D2 increased in isografts, and D3 increased in allografts. D3 was distinguished by high expression of migration-associated genes Ccr7 and Fscn1, which represented as DCs moving toward draining lymph nodes. Compared to the other two clusters of DCs, D3 has stronger phagocytosis ability and antigen presentation and process, especially in allografts.

Higher level of T cell activation was found in AR

2,604 T cells were clustered into 3 clusters. T1 increased in the Allo group compared with the other two groups. T1 could be divided into 2 subclusters: Cd4+T1and Cd8a+T1. Cd8⁺ T1 cells expressed Xcl1 and Ccl4. These genes are associate with recruiting DCs. Cd8⁺ T1 cells expressed Ldha. Ldha was responsible for increasing the pro-inflammatory phenotype of T cells. Cd4⁺ T1 cells had increased expression of Hif1α and Got1 genes. Cd4⁺ T1 cells highly expressed Tnfrsf4, which is a gene that functioned as a T cell co-stimulatory molecule. Interestingly, all the above six-highlighted genes had the highest expression in allografts. T2 contained T cells demonstrating higher scores of interferon-stimulated genes (ISGs), including Ifitm3, Irf7, and Gbp2. T2 cells were more abundant in normal hearts. T3 was characterized by the expression of Trdv4, Tcrg-V6, and Il17a, which were hallmarks of γδ T cells.

NK cells had higher cytotoxic effects in AR

We detected 1 cluster of 1,055 NK cells. All three groups had similar proportions of NK cells, around 2%. NK cells expressed more granular enzyme genes in allografts, including Gzma and Gzmb. Among the three groups, NK cells in allografts had the highest activated levels. Gene expression network analysis based on SCENIC analysis revealed Irf7 might play a central role in driving transcriptional activity of NK cells in allografts.

Hif1α regulated the degree of AR

Increased glycolysis levels of monocytes/macrophages were observed in allografts, but not in isografts. Aerobic glycolysis is the hallmark of metabolic switch in activated immune cells, which under the control of Hif1α. Hif1α was expressed highly in M1, M3, and T1 cells, which were all the main effector cells in AR. Similar to the elevated level of glycolysis, Hif1α was only obvious up-regulated in the Allo group. Immunofluorescence staining showed that HIF1A was highly expressed in allografts, especially in macrophages (HIF1A and CD68 positive).

To investigate the role of Hif1α in AR, we treated the AR mice model with the HIF1A inhibitor LW6. The intraperitoneal injection of LW6 into AR mice for 7 days attenuated leukocyte accumulation in allografts (No treat vs. LW6 treat, 3.6 ± 0.3 vs. 2.6 ± 0.2, p=0.019). The expression of HIF1A was inhibited by LW6 (Ctrl vs. No treat vs. LW6 treat, 8.3 ± 1.9 vs. 191.7 ± 55.7 vs. 30.9 ± 13.5, p=0.001). We further validated whether HIF1A is expressed in human AR heart specimens. Heart samples from patients with secondary transplantation due to AR were collected, compared with those from healthy controls. The number of HIF1A⁺ macrophages (CD68 positive) in patients with AR was increased.

Inhibition of HIF1A can restrict the pro-inflammatory behavior of macrophages

Another scRNA-seq was performed on the total cardiac immune cell population (CD45⁺) from the LW6 treat group and the no treat group. After quality control, we obtained 31,852 cells, in which LW6 treat group 16,197 cells, no treat group 15,655 cells. A special cluster of macrophages was identified, although belonging to the same cluster in the UMAP, a clear boundary existed between the LW6 treat group and the no treat group. The proportion of this cluster also increased after treatment. This cluster of macrophages may reflect the transcriptome changes responding to LW6 treatment, we defined it as LW6-associated macrophages. This suggested that LW6 treatment simultaneously reduced T cell infiltration proportionately, with unbiased effects on CD4 and CD8 T cells. Therefore, LW6 treatment may reduce the infiltration of T cells by reducing the antigen presenting ability and pro-inflammatory ability of macrophages.

Conclusion:

we described the immune microenvironment in cardiac tissues from the AR model by using scRNA-seq. HIF1A could promote the occurrence of AR by regulating the activation of macrophages. We proposed HIF1A may be a promising target in the therapy of AR but need precise control. Our study also provided a set of cell atlas for further functional research in AR.

Keywords: heart transplantation, acute rejection, Hif1α, immune metabolism.

Part 2. Model of Cardiac xenotransplantation based on wild-type pig donor and immunological rejection

Background: Cardiac xenotransplantation will be the direction of heart transplantation due to the shortage of donor organs. Although some great breakthrough has been made to increase the survival time after cardiac xenotransplantation, the fundamental immunological mechanism remained unclear. This projection aimed to investigate deeply immune rejection mechanism involved in the cardiac xenotransplantation through the advanced technology, such as single-cell RNA sequencing and to evaluate the feasibility of current gene editing strategies in combination with single cell sequencing data.

Methods: In this experiment, wild-type Bama mini pig and adult rhesus monkey were used as donor and recipient, respectively. The model was established by the intraperitoneal heterotopic transplantation, and the donor heart samples (n=2) and cardiac vessel samples (n=2) were obtained. Besides, the similar samples was also acquired from normal wild-type mini pigs as control. No-myocardial cells were isolated for single cell RNA sequence, cardiac proteins was extracted for Tandem Mass Tags proteomics and cytokines level in plasma was also detected by xMAP based on Luminex platform.

Results:

scRNA-seq landscape of hyperacute rejection after cardiac xenotransplantation

The heterotopic cardiac xenotransplants (n=2) were established for macaques with wild-type pig heart donors. The hyperacute rejection responses were present, including complement and coagulation cascade activation, endothelial activation and injury, and graft dysfunction (hyperacute rejection) within minutes or hours. Then cells from donor hearts and anastomotic vessels (pulmonary artery and aorta) were isolated for scRNA-seq, and cells from pigs without xenotransplantation (XT) were used as controls. There were 87092 cells detected in total, and after quality control, 86708 cells remained. Among these, 47484 cells were from XT group and 39224 cells were from control group. During cardiac xenotransplantation, macaque immune cells infiltrated into the donor heart. Therefore, we used a combined genome reference of pig and macaque for read alignment and gene expression quantification of scRNA-seq data. We first identified species origins of detected cells, and found that majority of cells were from pigs. The frequency of macaca immune cells infiltration was common into heart than anastomotic vessels.

Slight transcriptional changes appeared in pig non-immunce cells during xenotransplantation

The majority of pig cells were non-immune cells (more than 90%), which were identified with the expression of CD45. The proportions of non-immune cells of sus in hearts and vessels were similar. 75,510 non-immune cells could be clustered into 8 clusters with the classical marker genes expression, including fibroblasts (FBs), smooth muscle cells (SMCs), endothelial cells (ECs), Neurons (2 clusters), lymphatic endothelial cells (LECs), Dividing cells and cardiomyocytes (CMs). Considering the proportion of each cell cluster in different groups, the difference of cell proportion was almost similar except FB and SMC. The proportion of FB was highest in rejected heart, and the proportion of FB in heart and vessels were both higher in xenotransplantation group than in control group. Meanwhile, the proportion of SMC was lower in XT group than in control group.

Macrophage and CD8⁺ T cells underwent large proportional changes during xenotransplantation

8,256 pig immune cells were clustered into 13 clusters. The marker genes were used to identify the classical types, including macrophages, T cells, DCs, NK cells, B/plasm cells and so on. Macrophages and T cells comprised most of pig immune cells. B cells were characteristic with the expression of SOD1 and PLAC8; plasma cells were characteristic with the expression of JHCAIN and CD79B; CD8+ T cells were characteristic with the expression of CD3E, TRGC1, CD8A and CD5; NK cells were characteristic with the expression of KLRB1, KLRD1, KLRC1 and GNLY; macrophage cluster 1 was characteristic with the expression of CCL2, HMOX1 and CCL8. The proportions of CD8+ T cells and NK cells decreased in xenotranplanted hearts compared with controls, while the percentage of macrophage cluster 1 increased in xenotranplanted hearts. Enrichment analysis showed that macrophage cluster 1 was responsible for defense response, which might be triggered by injury and recognition of foreign antigens from macaques. CD8+ T cells and NK cells displayed high migration activity, so we reasoned that they migrated out of hearts and thus had decreased proportions. Differential gene expression analysis between xenotransplantation group and control group suggested that there were slight differences between them. HMOX1 was the hallmark gene which was expressed higher in xenotransplantation group.

Xenoantigens expression in pig cells

GGTA1 was expressed in highest percentage of cells (16.35%), followed by CMAH (3.20%) and B4GALNT2 (0.13%). Next, we analyzed the expression of three antigen genes in each cell types. FBs, SMCs, ECs and CMs were four largest populations of non-immune cells in hearts. Among them, ECs had highest proportion of these antigens. In immune cells, macrophages, DCs, B cells and Plasma cells had high proportion of these antigens. These immune cells received little interest in previous studies, maybe because they were rare cell populations and played no structural roles. In addition, we noticed that diving cells generally had high proportion of these antigens than other cells, which might suggest that dividing cells are more prone to transplantation rejection. We observed no large discrepancy in antigen expression between xenotransplantation group and control group.

T cells and monocytes are major infiltrating macaque immune cells

The chemokines, such as MCP-1, MIP-1β, Eotaxin and MIG, were increased in serum after XT. And there were 1,072 macaque cells detected. CD8+ T cells, monocytes and CD4+ T cells were three cell types with most cells. Similar infiltrating cell types appeared in hearts and vessels, but less cells infiltrated into vessels than hearts. ECs and FBs were almost only present in vessels, which might be from macaque anastomotic vessels. In the content of infiltrating cells, CD8+ T cells were the largest cell population in hearts, followed by monocytes and CD4+ T cells, while monocytes were the largest population in vessels. CD8+ T cells were characteristic with the expression of NKG7, GZMB, CD3E and CST7; CD4+ T cells were characteristic with the expression of CD3E, SPOCK2, CD7, CCR7, LTB and ICOS; monocytes were characteristic with the expression of S100A8, S100A9, THBS1 and LYZ. The functional enrichment analysis suggested CD8+ T cells mainly functioned in cytotoxicity, so they might participate in kill donor cells during hyperacute rejection; CD4+ T cells played roles in multiple peptide synthesis processes as helpers; monocytes could migrate into donor hearts and promote inflammation. The ligand-receptor interaction between macaque cells and pig cells were predicted, suggesting that macaque cells could potentially interact with pig cells. As ECs were the major targets during the hyperacute rejection process, we focused on the potential interaction between macaque cells and pig ECs. We found that the interaction between DCs and ECs was most significant. DCs interacted with ECs through the pair of TGFB1-CAV1 and THBS1-CD36; CD8+ T cells interacted with ECs through the pair of TGFB1 and CAV1/ENG; CD4+ T cells interacted with ECs through the pair of TGFB1-CAV1/ENG and CGA-RAMP2. Overall, CAV1 could interact with several ligands from all pig immune cells, and might be the most important EC receptor in xenotransplantation rejection.

The proteins changes was significant, and the differentially expressed proteins were involved in the IL-2 production pathway, neutrophils infiltration and thrombosis and spliceosome pathways. Finally, the gene modification strategy of xeno-donor pigs was re-evaluated, and new suggestions were proposed to improve the current gene modification strategy based on single cell RNA sequence data of immune rejection, so as to construct the best gene edited pigs.

Conclusion:

This study provides a method for the xenotransplantation model establishment, reveals the changes of pig and monkey cells during the process of xenotransplantation immune rejection by using single-cell technology, TMT proteomics etc, and puts forward new suggestions on the current gene modification based on the immune rejection mechanism of xenotransplantation, so as to promote the improvement of donor pig for xenotransplantation.

Keywords: Cardiac xenotransplantation, single cell RNA sequencing, wild-type pig donor, proteomics, novel gene modification pig.

 

 

Part 3. Single-cell RNA sequencing to dissect the immunological network of autoimmune myocarditis

Background:

Myocarditis can develop into dilated cardiomyopathy (DCM), which may require heart transplantation (HTx). MCD is characterized by a complex inflammatory response in the heart, and the clinical and histological manifestations are the result of aberrant immune processes. The pathological progression of myocarditis consists of 3 phases: acute inflammatory, subacute inflammatory, and chronic myopathy. Although some advances have been made regarding the mechanism of myocarditis, the complexity of heart cell infiltration remains poorly understood, partially due to the historical lack of methods designed to study these populations at a single-cell resolution. Moreover, the contributions of certain immune cell types to the inflammatory response to myocarditis are unknown, which limits the development of therapeutic drugs. The immunological network of myocarditis phases remains unknown. This study aimed to investigate the immunological network during the transition from myocarditis to cardiomyopathy and to identify the genes contributing to the inflammatory response to myocarditis.

Methods: Mice were treated with myosin heavy-chain-α peptides to generate an experimental autoimmune myocarditis (EAM) model. We performed single-cell RNA sequencing (scRNA-seq) analysis of Cd45⁺ cells extracted from mouse hearts during different EAM phases, including normal control, acute inflammatory, subacute inflammatory and myopathy phases. Human heart tissues were collected from the surgically removed hearts of patients who had undergone HTx.

Results:

scRNA-seq and cell typing of myocarditis at different time points

A model of EAM at different phases, including the normal control phase (day 0), acute inflammatory phase (day 14), subacute inflammatory phase (day 21) and myopathy phase (day 60), was established. The left ventricle end-diastolic dimension and left ventricle end-stage volume were the largest in the myopathy phase. However, the left ventricular ejection fraction (LVEF) and fractional shortening (FS) were the lowest in the myopathy phase (LVEF: day 0, 73.00 ± 7.61%; day 14, 66.85 ± 7.10%; day 21, 62.17 ± 10.63%; day 60, 58.08 ± 9.10%, p=0.0019. FS: day 0, 41.39 ± 6.41%; day 14, 36.47 ± 5.24%; day 21, 33.08 ± 8.16; and day 60, 30.36 ± 5.98, p=0.0020). Inflammatory cell infiltration in the heart tissue was most severe on day 14, decreasing on day 21 and further decreasing on day 60. The harvested cardiac Cd45⁺ cells from different phases were sequenced on a 10x Genomics platform. 34,655 cells were included in the subsequent analysis after quality filtering was performed. Subsequently, we identified macrophages (9 cell clusters), neutrophils (3 cell clusters), T cells (5 cell clusters), natural killer (NK) cells (1 cell cluster), B cells (2 cell clusters), dendritic cells (DCs; 3 cell clusters), innate lymphoid cells (ILCs; 1 cell cluster), cardiomyocytes (CMs, 1 cell cluster) and endothelial cells (Endos, 1 cell cluster). Macrophages constituted the major populations of immune cells at each phase (range: 61.06-72.88%, average: 65.03%), followed by B cells on day 0, neutrophils on both days 14 and 21, and T cells on day 60.

Myocarditis-associated macrophages are characterized by the expression of Hif1a, and different macrophage clusters contribute to the different EAM processes

We detected 22,348 macrophages as the largest cell population, which were clustered into 9 clusters and contributed to various EAM phases differently. We identified marker genes for each cluster, such as Cxcl9 for cluster 2, Ccl8 for cluster 3, Vcan for cluster 6 and Vsig4 for cluster 8. The percentage of cluster 2, the major macrophage cluster of the acute inflammatory phase, peaked on day 14 and was consistent with the extent of the inflammatory response. Nos2, Arg1 and Ass1 were differentially expressed in cluster 2, and the functions of cluster 2 included the nitric oxide (NO) biosynthesis process, phagosome and antigen processing and presentation of endogenous peptide antigens via major histocompatibility complex (MHC) class I molecules. Ccl8 was found to be differentially expressed in cluster 3, and cluster 3 was the hallmark macrophage population of the subacute inflammatory phase, with its percentage peaking on day 21. The functions of cluster 3 included antigen processing and presentation of exogenous peptide antigens via MHC class II molecules, cell chemotaxis, and mononuclear cell migration. Tnf, Il-10 and Vsig4 were differentially expressed in cluster 8, for which the percentage was lowest on day 14 and then gradually increasing as EAM progressed. Cluster 8 was shown to play a role in the positive regulation of the defense response. Cluster 7 contained M2-polarized macrophages expressing Clec10a and differentially expressed Tnip3. The cluster 7 percentage peaked on day 60 of EAM progression, and cluster 7 likely inhibited inflammation at the myopathy phase. SCENIC identified Bhlhe40, Nfil3, Mtf1, Esrra, Rnux3 and Hif1a as candidate TFs underlying the differences in gene expression in macrophage cluster 2 and Crem, Irf7, Bach2, Hif1a and Rxra as candidate TFs underlying the differences in gene expression in macrophage cluster 3.

The neutrophil cluster releases Il-1 to participate in the progression of EAM

We detected 4,299 neutrophils (more than 86% from the inflammatory phases) that were assembled into 3 clusters. Increased Il-1a expression was observed in clusters 1 and 3, whereas Il-1b was expressed at high levels in all three clusters, particularly in neutrophil cluster 1. Cluster 1 formed the major proportion of neutrophil populations on day 14, and this cluster produced cytokines such as IL-6 and TNF. Furthermore, the functions of cluster 1 included regulation of the defense response, neutrophil degranulation, and the inflammatory response. Neutrophil cluster 1 (Il1a⁺ neutrophils) is a proinflammatory cluster in the early EAM inflammatory phase. Cluster 2 constituted the major neutrophil population (greater than 80%) at myopathy phase and highly expressed the profibrotic molecule Mmp8. Cluster 2 functioned in the positive regulation of the MAPK cascade and actin cytoskeleton organization, neutrophil cluster 2 (Ngp⁺ Mmp8⁺ neutrophils) is the main neutrophil cluster involved in the myopathy phase of EAM.

Il-17a⁺ T cell clusters may play an essential role in the dynamic immune environment of EAM

A total of 3,805 T cells that were clustered into 5 clusters. Tnfsf8 and Cxcr6 were marker genes of cluster 2, which also expressed Cd4, Il17a, Rora and Tnfsf11. The functions of cluster 2 included T cell activation, Th17 cell differentiation, Th17 type immune response and Th17 cell lineage commitment. Based on these results, T cell cluster 2 was comprised of Th17 cells. The percentage of Th17 cells among total immune cells peaked on day 14, and its change trend was consistent with that of the percentages of immune cells in the four phases. T cell cluster 4 was characterized by the expression of Il23r, Il-17a, Tcg-v2 and other T cell receptor genes and thus represented γδ T cells. Moreover, according to gene enrichment analysis, T cell cluster 5 was also defined as γδ T cells. T cell cluster 4 expressed Il-17a (Il-17a⁺ γδ T cells), whereas T cell cluster 5 did not (Il-17a^- γδ T cells). In addition, T cell cluster 4 (Il-17a⁺ γδ T cells) was the major cell type on day 60. T cell receptor (TCR) sequencing data showed that TCR clones were mainly detected in cluster 2 (Th17 cells) on day 21. Genes regulated by Th17 cell-specific TFs (Hif1a, Rora and Nfkb1) and by Xbp1 were upregulated in cluster 2 (Th17 cells). Genes regulated by Rara, Atf6, Fosb and Rora.1 and apoptotic genes regulated by Rarg were upregulated in cluster 4 (Il-17a⁺ γδ T cells).

B cells play roles in antigen presentation and antibody production in EAM

We detected 1,459 B cells that formed 2 clusters. B cell cluster 1 expressed Ighd at higher levels, whereas B cell cluster 2 expressed Igkc, Iglc1, Iglc2, Iglc3 and Ighj3 at higher levels. B cell cluster 1 was characterized by Ms4a1 (Cd20) and Cd19 expression. Cluster 1 was the major population (ranging from 92.2-98.5%) of B cells at the different EAM phases. In addition, B cell receptor (BCR) clones were detected during EAM progression and were found to be mainly present in cluster 1 on day 60. Cluster 2 was marked by Jchain and Eaf2 expression. Among all B cell populations, the cluster 2 population increased with the progression of EAM. Genes regulated by Elk3, Max, Irf2, Ikzf1, Spib, Pax5 and Sox5 were upregulated in B cell cluster 1. Importantly, Irf2 was a transcriptional activator of many key components of the MHC I antigen presentation. Together, these results show that cluster 1 (immature B cells) participates in antigen processing and presentation. Genes regulated by plasma-specific TFs (Prdm1, Atf6, Xbp1 and Creb3) were upregulated in cluster 2. Thus, B cell cluster 2 (plasma cells) produces antibodies to contribute into the autoimmunity of EAM.

DCs in EAM progression

A total of 1,077 DCs were detected and then separated into 3 clusters with different phase distributions. DC cluster 1 (Cst3⁺Sept3⁺ DCs) was the major DC population among all EAM phases, while the percentage of DC clusters 2 (Ccr9⁺Klk1⁺ DCs) and 3 (Il12b⁺Cacnb3⁺ DCs) increased at inflammatory phases. Interestingly, DC cluster 3 was present at low levels in the immune cell population of the normal heart (less than1‰), and its proportion increased substantially when the inflammatory response occurred but decreased as EAM progressed. DC cluster 3 was shown to be involved in antigen processing and presentation, myeloid cell differentiation and the production of cytokines such as TNF. SCENIC revealed that genes regulated by Runx1 and Irf5 were upregulated in DC cluster 1. Furthermore, genes regulated by Ehf, Nfkb1, Stat1, Rel, Nfkb2, and Stat4 were expressed at high levels in DC cluster 3. The expression of Rel, an NF-κB subunit, in DC cluster 3 promoted Il-23 expression to regulate the Th17 response. Taken together, these results indicate that DC cluster 3 may participate in the inflammation of EAM by activating the NF-κB pathway to regulate the Th17 response.

Correlations between immune cell subtypes and EAM phases suggest that Hif1a contributes to the inflammation observed in EAM

Some clusters, such as macrophage cluster 2 and Th17 cells, expressed inflammatory genes at high levels on day 14. Hif1a was shared by macrophage cluster 2 and Th17 cells and differentially expressed in the cell clusters. Furthermore, the expression level of Hif1a correlated with the cell cluster inflammatory score (Pearson’s r=0.69, p=0.001). Similarly, Hif1a was expressed at the highest levels on day 14, followed by day 21, day 60 and day 0. Together, these results suggest that Hif1a may activate the immune cell types associated with the acute EAM phase (macrophage cluster 2 and Th17 cells) to contribute to the pathogenesis of EAM.We established another EAM model (n=50 mice) to investigate the therapeutic effects of PX-478 on EAM at the different phases. These EAM mice were divided into groups covering three phases (day 14, day 21 and day 60), and PX-478 was injected into each EAM mouse for 7 days before the mice were sacrificed. Strikingly, the intraperitoneal injection of PX-478 into EAM mice attenuated leukocyte accumulation in the hearts and ameliorated inflammation on day 14. Furthermore, the expression of Hif1a was inhibited by PX-478. As Hif1a is primarily expressed in macrophages and T cells, PX-478 reduced the infiltration of T cells and macrophages on day 14. In addition, sensitized animals treated with PX-478 at other phases, including days 21 and 60, also had significantly reduced inflammatory responses. Taken together, these results indicate that Hif1a contributes to the inflammation observed in EAM, and its inhibitor, PX-478, inhibits inflammation at different phases. We found that compared with those in the DCM (n=5) and normal controls (n=5), the number of HIF1A⁺ cells per mm² in autoimmune myocarditis patient (n=5) samples was significantly increased (4,106.0 ± 423.6 vs. 605.6 ± 213.7 cells per mm², p<0.0001 and 4,106.0 ± 423.6 vs. 821.6 ± 223.0 cells per mm², p<0.0001, respectively). Finally, some T cells and macrophages and a few cardiomyocytes were positive for HIF1A in autoimmune myocarditis patient heart tissues. Together, these data demonstrate the clinical relevance of HIF1A to autoimmune myocarditis patients and may serve as a novel therapeutic target for autoimmune myocarditis patients.

Conclusions:

These data described the changes in the immunological environment of cardiac tissues from an EAM model at different phases using scRNA-seq. Some cell clusters correlated with the extent of the inflammatory response in EAM. Furthermore, the TF Hif1a is a potential candidate target for the treatment of autoimmune myocarditis.

 

Keywords: experimental autoimmune myocarditis, immunological network, single-cell sequencing, Hif1a, PX-478